Matches in SemOpenAlex for { <https://semopenalex.org/work/W3039185146> ?p ?o ?g. }
- W3039185146 abstract "Answering questions that involve multi-step reasoning requires decomposing them and using the answers of intermediate steps to reach the final answer. However, state-of-the-art models in grounded question answering often do not explicitly perform decomposition, leading to difficulties in generalization to out-of-distribution examples. In this work, we propose a model that computes a representation and denotation for all question spans in a bottom-up, compositional manner using a CKY-style parser. Our model induces latent trees, driven by end-to-end (the answer) supervision only. We show that this inductive bias towards tree structures dramatically improves systematic generalization to out-of-distribution examples, compared to strong baselines on an arithmetic expressions benchmark as well as on CLOSURE, a dataset that focuses on systematic generalization for grounded question answering. On this challenging dataset, our model reaches an accuracy of 96.1%, significantly higher than prior models that almost perfectly solve the task on a random, in-distribution split." @default.
- W3039185146 created "2020-07-10" @default.
- W3039185146 creator A5020117260 @default.
- W3039185146 creator A5035088083 @default.
- W3039185146 creator A5045872048 @default.
- W3039185146 creator A5051169643 @default.
- W3039185146 date "2020-07-01" @default.
- W3039185146 modified "2023-10-01" @default.
- W3039185146 title "Latent Compositional Representations Improve Systematic Generalization in Grounded Question Answering" @default.
- W3039185146 cites W1496189301 @default.
- W3039185146 cites W1502293651 @default.
- W3039185146 cites W1581407678 @default.
- W3039185146 cites W1970961429 @default.
- W3039185146 cites W2064675550 @default.
- W3039185146 cites W2117539524 @default.
- W3039185146 cites W2118373646 @default.
- W3039185146 cites W2152907450 @default.
- W3039185146 cites W2161002933 @default.
- W3039185146 cites W2194775991 @default.
- W3039185146 cites W2250385622 @default.
- W3039185146 cites W2250539671 @default.
- W3039185146 cites W2561715562 @default.
- W3039185146 cites W2613718673 @default.
- W3039185146 cites W2619818172 @default.
- W3039185146 cites W2745461083 @default.
- W3039185146 cites W2798663534 @default.
- W3039185146 cites W2883104598 @default.
- W3039185146 cites W2887020936 @default.
- W3039185146 cites W2889830342 @default.
- W3039185146 cites W2902442580 @default.
- W3039185146 cites W2932376173 @default.
- W3039185146 cites W2951328433 @default.
- W3039185146 cites W2953388933 @default.
- W3039185146 cites W2962716332 @default.
- W3039185146 cites W2963187627 @default.
- W3039185146 cites W2963223524 @default.
- W3039185146 cites W2963267799 @default.
- W3039185146 cites W2963305465 @default.
- W3039185146 cites W2963341956 @default.
- W3039185146 cites W2963403868 @default.
- W3039185146 cites W2963738360 @default.
- W3039185146 cites W2963921132 @default.
- W3039185146 cites W2964118342 @default.
- W3039185146 cites W2970231061 @default.
- W3039185146 cites W2975501350 @default.
- W3039185146 cites W2995479498 @default.
- W3039185146 cites W3013503438 @default.
- W3039185146 cites W3026376263 @default.
- W3039185146 cites W3034873522 @default.
- W3039185146 cites W3035267217 @default.
- W3039185146 cites W3035428952 @default.
- W3039185146 cites W3037932933 @default.
- W3039185146 cites W3098741467 @default.
- W3039185146 doi "https://doi.org/10.48550/arxiv.2007.00266" @default.
- W3039185146 hasPublicationYear "2020" @default.
- W3039185146 type Work @default.
- W3039185146 sameAs 3039185146 @default.
- W3039185146 citedByCount "4" @default.
- W3039185146 countsByYear W30391851462021 @default.
- W3039185146 crossrefType "posted-content" @default.
- W3039185146 hasAuthorship W3039185146A5020117260 @default.
- W3039185146 hasAuthorship W3039185146A5035088083 @default.
- W3039185146 hasAuthorship W3039185146A5045872048 @default.
- W3039185146 hasAuthorship W3039185146A5051169643 @default.
- W3039185146 hasBestOaLocation W30391851461 @default.
- W3039185146 hasConcept C113174947 @default.
- W3039185146 hasConcept C119857082 @default.
- W3039185146 hasConcept C134306372 @default.
- W3039185146 hasConcept C138885662 @default.
- W3039185146 hasConcept C139997677 @default.
- W3039185146 hasConcept C149106765 @default.
- W3039185146 hasConcept C154945302 @default.
- W3039185146 hasConcept C162324750 @default.
- W3039185146 hasConcept C177148314 @default.
- W3039185146 hasConcept C17744445 @default.
- W3039185146 hasConcept C186644900 @default.
- W3039185146 hasConcept C187736073 @default.
- W3039185146 hasConcept C199539241 @default.
- W3039185146 hasConcept C204321447 @default.
- W3039185146 hasConcept C2776359362 @default.
- W3039185146 hasConcept C2780451532 @default.
- W3039185146 hasConcept C33923547 @default.
- W3039185146 hasConcept C41008148 @default.
- W3039185146 hasConcept C41895202 @default.
- W3039185146 hasConcept C44291984 @default.
- W3039185146 hasConcept C94625758 @default.
- W3039185146 hasConceptScore W3039185146C113174947 @default.
- W3039185146 hasConceptScore W3039185146C119857082 @default.
- W3039185146 hasConceptScore W3039185146C134306372 @default.
- W3039185146 hasConceptScore W3039185146C138885662 @default.
- W3039185146 hasConceptScore W3039185146C139997677 @default.
- W3039185146 hasConceptScore W3039185146C149106765 @default.
- W3039185146 hasConceptScore W3039185146C154945302 @default.
- W3039185146 hasConceptScore W3039185146C162324750 @default.
- W3039185146 hasConceptScore W3039185146C177148314 @default.
- W3039185146 hasConceptScore W3039185146C17744445 @default.
- W3039185146 hasConceptScore W3039185146C186644900 @default.
- W3039185146 hasConceptScore W3039185146C187736073 @default.
- W3039185146 hasConceptScore W3039185146C199539241 @default.
- W3039185146 hasConceptScore W3039185146C204321447 @default.